What does a reaction test actually measure? Data vs. intuition
After building Reaction Dot, I tested myself for a week, averaging 248ms — thinking "my reaction is average." Then I read some neuroscience papers. Realized several things that changed how I think about "reaction speed." Sharing for others who care about what's behind the numbers.
Reaction time isn't one number
Most tests give you a total — like 230ms. But physiologically, this is composed of at least 3 independent parts:
1. Sensory transmission (~30-50ms). Screen turns green → photons hit your retina → nerve impulse travels via the optic nerve to visual cortex. Mostly determined by physiology; nearly impossible to improve.
2. Central processing (~100-150ms). Visual cortex recognizes "it lit up" → decide to press the button. The biggest segment, most affected by cognitive state (fatigue, stress, caffeine all hit here).
3. Motor output (~60-100ms). Motor cortex issues command → travels down the spinal cord to finger muscles → finger clicks → mouse/keyboard fires → browser captures event. Most of this is muscle contraction itself.
30 + 120 + 80 = 230ms. Typical "simple reaction" score. Athletes, esports pros, and youth advantage can push to 180-200ms but rarely break 150ms — physiological transmission has a hard floor.
Common misconceptions
After a week of playing Reaction Dot and chatting with friends about results, several myths are widespread:
Myth 1: "My reaction time is 250ms, same as yesterday."
Truth: reaction times fluctuate hugely. The same person — post-coffee + morning + 8h sleep, vs. post-beer + 2am + 5h sleep — can differ by 50ms. A single test result is nearly meaningless; you need a week of data to see a curve.
Myth 2: "Esports pros must have fast reactions."
Truth: Research (Mann et al. 2019 on esports) found top esports pros' simple reaction time is similar to casuals — typical 230-260ms vs. pros 200-230ms. Small gap. But choice reaction and prediction differ hugely — pros predict positions before the ball appears.
Myth 3: "I'm slow because my brain is slow."
Truth: most "slow reaction" is actually slow motor. From signal to decision, central processing is similar across people; what differs is "execution delay after decision." This delay depends on muscle state, input device (mechanical keyboard vs. touchscreen), posture.
Myth 4: "Daily practice makes me faster."
Truth: simple reaction has a physiological floor. Long-term practice can shave 30-50ms but plateaus around 180ms. Continued practice only maintains. Age effect is larger: 25-year-olds beat 40-year-olds by 20-40ms.
Why Stroop is so slow
Reaction Dot's Stroop mode might surprise you. Simple reaction 250ms, choice reaction 350ms, Stroop reaction 500-700ms. More than double.
Reason: Stroop tests "inhibitory control" — a much higher cognitive capability than reaction.
When you see a red-colored word "GREEN," your brain receives two signals simultaneously:
- Visual: "color is red"
- Language (auto-activated): "the word is 'green'"
Normally, the language system responds faster (decades of reading have automated semantic decoding). So your brain must "suppress" the language signal and pick the visual signal — this suppression needs an extra 100-300ms of prefrontal cortex activity.
This delay difference ("Stroop interference") is the gold standard for measuring executive control. Clinically used for assessing ADHD, Alzheimer's, prefrontal damage. Reaction Dot's Stroop data isn't diagnostic but reflects your "current focus" well.
On caffeine, alcohol, sleep
Weekly self-test trends (N=1, not rigorous):
Caffeine (before/after 2 coffees): simple reaction 15ms faster, Stroop 30-50ms faster. Consistent with Cambridge 2010 caffeine research (they found caffeine mainly improves "central processing speed," not "sensory transmission").
Sleep deprivation (5h vs 8h): simple reaction 25ms slower, Stroop 100-150ms slower. Consistent with "sleep deprivation mainly affects prefrontal cortex" consensus.
After 2 beers: simple reaction 50ms slower, Stroop 200ms+ slower, and "early click" frequency rose significantly — you click prematurely more often. Alcohol lowering inhibitory control is clearly visible.
Using Reaction Dot to self-test caffeine/sleep/alcohol effects is surprisingly intuitive.
True reaction speed is mostly determined by physiology (~200ms). The remaining 30-50ms optimization comes from focus, sleep, caffeine. Practice all you want — you can't become superhuman.
Practical advice for gamers
If you play FPS or action games and care about "improving reactions":
1. Sleep well. More important than any "reaction training." 8h vs 5h, all your reaction data differs by 50ms+.
2. Moderate caffeine, not excessive. 200mg adds 15-30ms; 400mg+ creates hand tremor and worse judgment.
3. No drinking before gaming. Even one beer: Stroop-class inhibitory control drops noticeably.
4. Train "prediction," not "reaction." Top players aren't fast-reacting; they're accurate-predicting. They expect an enemy to appear at a corner → pre-aim there → no reaction needed when enemy shows. This "prediction training" is what truly separates experts from casuals.
5. Input devices matter. High-polling-rate mouse (1000Hz vs 125Hz), low-latency monitor (1ms vs 5ms), wired vs wireless — these cumulatively save 10-30ms, more reliable than biological optimization.
Closing
I used to think "reaction speed" was a single number. This week's testing made me realize it's a composite indicator from physiology, cognition, and equipment. In-game reaction tests measure "simple reaction" and "choice reaction" — neither has much impact on real life. What truly determines life and work performance is "inhibitory control" and "prediction" — Stroop and real practice are the mirrors for those.
So next time Reaction Dot gives you 250ms, don't be anxious or proud. It's just the tip of the iceberg.
Max is BverGame's co-operator. References include Stroop's original 1935 paper, Mann et al. 2019 on esports reaction times, and Cambridge's Brice & Smith 2010 caffeine research. Personal sample size is very small; consult professionals for clinical relevance.